A Madison-based team has stripped liquid chromatography out of peptide mapping, the workhorse test drug developers use to check whether antibody therapeutics have degraded. The new method, described in a bioRxiv preprint and now spun out as the company CeleramAb, promises to turn a nine-day analysis into a matter of hours.

Peptide mapping is one of biopharma’s more thankless jobs. Every monoclonal antibody destined for the clinic must be enzymatically chopped into peptides, separated by liquid chromatography (LC), and fed into a mass spectrometer to check for oxidation, deamidation, clipping and other chemical scars that can blunt a drug’s effectiveness. It works, but it is slow, and demand for the test has long outstripped laboratory capacity.

A team led by Joshua Coon, metabolism investigator at the Morgridge Institute for Research and professor of biomolecular chemistry at the University of Wisconsin–Madison, believes it has found a way round the bottleneck. Writing in a preprint posted to bioRxiv on 16 May 2026, Coon and colleagues describe RaPiD-mAb-MS, a direct-infusion mass spectrometry method that removes the chromatography step entirely.

A test that has outgrown its own throughput

According to the paper’s authors, the technique “combines high-throughput plate-based sample preparation with direct infusion mass spectrometry analysis” and can analyse 96 samples in roughly 1.5 to 2 hours while routinely achieving greater than 95 per cent sequence coverage. Rather than separating peptides by retention time, samples are sprayed directly into an Orbitrap-based mass spectrometer using an automated nanoelectrospray platform, with high-resolution MS1 and targeted MS/MS scans used to detect and quantify modifications from mass alone.

“Companies today can do about a sample every hour, so people are doing at best 15 to 20 a day,” Coon said. “We can do 30 to 40 an hour, or 600 a day. So that’s pretty transformative.”

Benchmarked against industry’s gold standard

The claim is not merely theoretical. The group tested RaPiD-mAb-MS against conventional LC-MS across 28 unique antibodies and more than 2,000 samples, working with AbbVie, Genentech, Eli Lilly and Johnson & Johnson. In one head-to-head comparison, 275 forced-degradation NIST monoclonal antibody samples that had taken AbbVie eight days of LC-MS runtime plus a further day of data searching were reanalysed by the new method in around 12 hours, with the authors reporting “excellent agreement” between the two techniques for oxidation and deamidation measurements. A separate 1,152-sample cohort run at Johnson & Johnson, spanning four antibodies and dozens of buffer and stress conditions, was completed in roughly 35.5 hours of instrument time.

The method also picked up harder-to-detect features, including sequence variants at around 0.7 per cent occupancy, glycoform ratios, and drug-antibody conjugation sites on two antibody-drug conjugates from Genentech. Isoaspartate, a modification normally invisible to mass spectrometry without chromatographic separation, was made detectable through an added enzymatic labelling step using the enzyme PIMT.

From bench discovery to bootstrapped biotech

The underlying chemistry traces back to 2018, when the Coon lab first found a chromatography shortcut in shotgun proteomics. Alex Hebert, a scientist on that project, later moved into industry and saw the LC bottleneck firsthand before returning to help adapt the approach for antibody characterisation. Coon and UW–Madison chemistry professor Lloyd Smith, who previously co-founded Third Wave Technologies, have licensed the relevant intellectual property from the Wisconsin Alumni Research Foundation and built CeleramAb around it, deliberately avoiding outside investors.

“Bootstrapping allows us to remain highly agile and strictly focused on our customers’ needs rather than capital demands,” Smith said. “We operate incredibly lean, which has allowed us to rapidly iterate our technology alongside our top pharma partners and build a highly scalable commercial platform.”

Implications for drug development pipelines

Coon argues the gains extend beyond convenience. “Once they get familiar and comfortable with our process, we anticipate even more of these experiments being done,” he said. “That’s going to not just save the company money, but it’s going to accelerate their ability to move drugs through the pipeline. That’s going to help everybody.”

The authors are careful to note that conventional LC-MS will remain necessary for final regulatory quality control, given its entrenchment in FDA-approved methods. But for early discovery and process development, where stress-testing dozens of candidates or running full factorial experiments has previously been impractical, they suggest the technology “sets the stage for collection of massive datasets” that could eventually feed machine learning models predicting optimal antibody formulations and sequences.

Journal reference: Salome, A. Z., Morgenstern, M., Hebert, A. S., et al. (2026). Rapid peptide mapping of monoclonal antibodies with direct infusion mass spectrometry. bioRxiv. https://doi.org/10.64898/2026.05.14.725248